Skip to content

Write a power of a sum of exponentials that is one exponential as it - #2043

Closed
Rafael-SOWNet wants to merge 1 commit into
masterfrom
fold-a-power-of-a-sum-of-exponentials-into-one
Closed

Rafael-SOWNet wants to merge 1 commit into
masterfrom
fold-a-power-of-a-sum-of-exponentials-into-one

Conversation

@Rafael-SOWNet

Copy link
Copy Markdown
Member

Part of #233.

  • A power of a sum of exponentials that is one exponential, written as it. The parser writes cosh(u) + sinh(u) as (e^u + e^(-u))/2 + (e^u - e^(-u))/2, and nothing collected it. Rubi's 6.7.1 F^(a + b x) (cosh(c + d x) + sinh(c + d x))^n and its twin with the difference were declined, and (cosh(x) + sinh(x))^3 was answered expanded.
  • How it reads the sum. At the question asked, a power of a sum of exponentials of x, each a whole power of the first, w, is read as a Laurent polynomial in w. Where one coefficient is left, the sum is c w^p. (c w^p)^n is c^n w^(p n) for a positive w, which e to a real power is.
  • Two details. Expanded, the sum came back as w provided not w = 0, so it is read through Bare. The answer is inner-simplified, which takes the ln(e) the rule for F^u writes back to 1.
Integrand master 78cd31b this branch
6.7.1:1259 F^(a + b x) (cosh(c + d x) + sinh(c + d x))^n declined 160 characters
6.7.1:1260 F^(a + b x) (cosh(c + d x) - sinh(c + d x))^n declined 165 characters
(cosh(x) + sinh(x))^3 551 characters (787 on 2.5.0) e^(3 x)/3

Chain against master 78cd31b:

  • The pocket of every family-2 and family-6 row, 4,340 of them, went from 4,315 to 4,317 solved. The one "wrong" on each side is 6.1.5:94, the evaluator's.
  • Every family's sample is unchanged.
  • The two rows that moved, 1259 and 1260, run alone, went from 0 to 2 solved.
  • 9,855 rows were compared for allocation, and none grew past 1.3x and 300 MB.

The full unit suite passes locally: 16,553 passed, 0 failed. The BREAKING-CHANGES lengths were measured on each build. Rebased onto 5b85e6b.

🤖 Generated with Claude Code

https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura

The parser writes cosh(u) + sinh(u) as (e^u + e^(-u))/2 + (e^u - e^(-u))/2,
and nothing collected it: Rubi's 6.7.1 F^(a + b x) (cosh(c + d x) + sinh(c + d x))^n
and its twin with the difference were declined, and (cosh(x) + sinh(x))^3 was
answered in 551 characters, expanded. At the question asked, a power of a
sum of exponentials of x, each a whole power of the first, w, is read as a
Laurent polynomial in w, and where one coefficient is left it is c w^p:
(c w^p)^n is c^n w^(p n) for a positive w, which e to a real power is. The
integral of what it folds to is asked as the same question and checked.

Part of #233.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura
@Rafael-SOWNet

Copy link
Copy Markdown
Member Author

Closing as redundant: #2034, merged since this branch was cut from 78cd31b, already answers Rubi's 6.7.1:1259 and :1260 on master 49bf5d1, in 141 and 142 characters where this branch writes 160 and 165, measured on a build of each. What is left here is (cosh(x) + sinh(x))^3 in 19 characters where master writes 551, which is not a corpus row; it is noted for later rather than kept as its own PR. The branch stays on the remote.

🤖 Generated with Claude Code

@github-project-automation github-project-automation Bot moved this from Todo to Done in AngouriMath Oct 22, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

1 participant